What's Holding Women Back?

Is discrimination still a thing?

i don’t actually watch industry so you can’t yell at me if the image isn’t accurate. i’ll get around to it after interview with the vampire

On Sunday I visited my grandmother, which was fun until my family got stuck in traffic for three hours on our way back. Alongside reading two books because I forgot my phone at home and talking about when doctors used to prescribe leeches to people, the lunch table conversation moved to politics - particularly whether Americans would be capable of voting for a woman (Kamala Harris, obviously) or not. Something my grandmother mentioned is that she would like to vote based on qualifications and ideas, not gender, and brought up how the country’s worst-ever president was a completely unqualified woman (#Feminism #Herstory)¹, which triggered a long discussion about whether gender should be considered as a factor in politics or not.

Here’s where the conversation gets interesting: my mother said that she disagreed because obviously women face a lot of discrimination, and cited some personal examples. My grandmother, who is 86 years old, said she hadn’t felt personally discriminated, except on a handful of occasions, but brought up some interesting family history. Her grandfather, a businessman, refused to allowed my great grandmother to go to school to become a teacher, since she would have to travel too far away. The way this ban was enforced was simple: it was socially unacceptable for women to leave the house without wearing a hat, and my great great grandfather kept all the hats in the house locked. Therefore, his daughter did not become a teacher. However, she and her husband only had daughters, and both wanted the family soda company to stay in their hands, so my grandmother and her three sisters all attended university, which was, quite obviously, fairly rare for women born in the 1930s and 1940s. My grandmother travelled to Buenos Aires, the capital, and took a longer time than usual to finish her architecture degree because she was married and had children. She’s been working for the last 60 years or so, and all her sisters also completed degrees (two of them were accountants, an even less commonly female profession) and worked for most, or all, their adult lives. My mother also went to school in the City, and her family gave her some support while she completed her degree. And now I have my own degree and a nice good job.

The point of this blog post is, more or less, to settle a family argument: who was right about when women stopped being discriminated against?

Get back in the kitchen

Is there discrimination against women? Kind of. The starting point would obviously be the one very famous paper about orchestra auditions: using data from the audition records of eight major symphony orchestras from the 1950s until 1995 at the individual level, the musicians are characterized by gender and whether or not the audition was completely blind or not completely blind. The paper estimates a probability function that takes into account whether the audition (each round) was blind or not blind, as well as the time period (to allow for changes in social attitudes, for instance) and the specific orchestra, to remove potential selection effects. The “treatment group” are orchestras with fully blind auditions, and the control are orchestras without them. Importantly, there are also controls for each individual musician, to allow for less talented or experienced players to be penalized - while this might conceal other forms of sexism, it does remove direct bias from the data.

The results are quite clear: using blind auditions throughout the hiring process (which has multiple stages) increases the probability for women after controlling for individual factors, except for the semifinal round, which has certain particularities that make finding specific impacts difficult. One major factor to point out is that the results are not statistically significant, in general, but mainly due to the fact that success rates in auditions are vanishingly small in general, even for men - so, as a result, the effect of a 50% increase in likelihood of a woman advancing in interviews is very small.

Empirically, discrimination is what we call an unobservable - it can’t be measured. The fraction of the wage gap determined by variables that can’t be measured shrank from the 1960s to the 1990s, but didn’t really shrink any further. This could, in principle, mean that discrimination declined over time, which seems fairly common sense. To date, discrimination seems to play a small (though sometimes real) role in wages and pay for women. Women, for example, get punished for the same mistakes more harshly than their male peers. There are cases where women are discriminated against in particular ways, such as Muslim women being less likely to get called back for jobs, or older women being considered less qualified than younger women. And back in the 1960s, equal pay discrimination allowed women’s earnings to increase.

But while this can be said about the past, it can’t really be said about the present. Misogynistic attitudes and beliefs still exist, and there are many disparities between men and women, but it’s not really true that there’s widespread negative attitudes towards women working - in most developed and developed-adjacent countries, it’s generally agreed upon that women are capable of doing work as good as a man’s. Less enlightened societies, however, do have fairly extensive records of violence and exclusion against women - for example, a third of Egyptians told a survey that honour-killings of women should go unpunished. And in India, gender discrimination goes so far that baby boys are significantly more likely to be breastfed than baby girls.

Something that is kind of usually omitted from this conversation is sexual harassment, and while its actual prevalence is a matter of debate, it has become significantly less acceptable as of late. This is good, as sexual harassment is something of a tax on women’s ambition: people subject to sexual harassment are disproportionately likely to quit their job and take a different one, which on average pays less and is less professionally fulfilling. Additionally, it warps workplace functioning, since in order to avoid missing out on positive relationships with their superiors, individuals underreport sexual misconduct. And when people are aware of sexual misconduct against people of their own gender, they tend to avoid working at certain firms, but do basically nothing when it affects the opposite gender; in this sense, tolerance of workplace harassment can be understood as a form of gender discrimination, particularly since the vast majority of instances are against women. Fortunately, it does seem that sexual harassers do suffer professional consequences, since scientists with harassment allegations have lower citation counts. However, it is a double edged sword, since the social dynamics at play are complex, and accusing superiors may also disincentivize male-female mentorship out of fear of being falsely accused. In economics, for instance, after the MeToo movement (which was good, for what it’s worth), cooperation between male and female economists increased at the coauthor level, but decreased between senior men and junior women in ways that are associated with likelihood of being accused.

It’s a Kamalanominon

Why hasn’t the United States had a female President? Well, until 1920, women didn’t have the right to vote, and until the 1980s and 1990s there were very few women elected to high offices. If most Governors and Senators are men, then most presidential candidates will be men, since that is the usual pool from which they emerge. And not a lot of women from that pool have even run for president anyways; before Hillary Clinton, only 5 or 6 women even ran for President, and only two were Senators or “more”. So it wasn’t a deep pool until at least the 1990s!

Beyond the flaws or chances of each specific candidate, a major concern among voters (and my non-American family) is that Americans would not elect a woman over a man - but research has found that women candidates, after accounting for partisanship, candidate quality, and other exogenous effects, perform around as well as men do. Simply put, a solid and well qualified woman running for President should do as well as a man with similar qualifications. The problem isn’t really that women in politics are on average less charismatic, or less competent, or have worse luck - it’s that fewer women think they themselves are qualified to run for office.

This, of course, tracks with general perceptions that women are less ambitious than men: as part of a review of Sheryl Sandberg’s Lean In philosophy, I mentioned that there is clear evidence that women negotiate less at work, but it is driven by average personality differences to some extent, differences in information and expectations about expected pay relative to qualifications (which might be related to women valuing their own qualifications less highly), and selectivity choosing when to negotiate that leads to excessive caution.

This is generally in line with the research: for example, women tend to apply to jobs differently than men (the biggest reason is that women apply to jobs that are more family-friendly, and thus exchange lower pay for a higher ability to participate in family life). Women also tend to gravitate towards careers they find more personally fulfilling over careers that are more financially rewarding, resulting in higher female labor satisfaction but also lower expected earnings. I’ve written about the gender gay pap to some extent in the past, and the takeaway is mainly that it partly reflects career choice, but even within occupations it has some components resulting from women’s decisions to work fewer hours and take more time off, which is also related to women undertaking most of a family’s domestic labor.

Similarly, researchers from Germany found that women are less likely to move forward and apply for promotions, largely because of the way these positions are structured and because of lower trust in future performance. This isn’t exclusive to women: men are less likely to apply to work in female-dominated spaces unless specifically prompted to trust in their own ability. There are many reasons why management in firms is disproportionately male (a big one is that management is disproportionately older, and older senior workers are more male than younger ones), but having access to professional networks is very important, and in many cases these networks exclude women. One major and not especially applicable to anything else example is that dating a senior coworker increases your earnings - not because of selection effects, but rather, because of favoritism. This tracks with other research, such as papers finding that having more friends as a teenager increases future earnings.

No exit

Recently, there’s been a lot of talk about make-up on Twitter, after Derek Guy the menswear guy tweeted that women like to dress slutty and wear makeup not necessarily for the approval of men. A lot of people disagreed, and while I do think that pushing back on the naive pop choice feminism of “I do it for myself” is productive, I also fundamentally believe that people can’t really make decisions about socially valued actions in ways that are completely independent of the social value. Women might dress in certain ways because they sincerely believe it looks good, but “what looks good” is not some a priori synthetic truth of the universe. It’s a social construction. I don’t wear make-up on my day to day life (I have sensitive skin and honestly just don’t feel like it), but I still realize that there’s kind of an unspoken mandate for that kind of thing. Like, if you don’t shave your armpits there’s nobody who is going to punch you in the face, but it does feel like you’re “breaking the rules of attractiveness” in some way.

In his play No Exit, Jean Paul Sartre has a famous line that “hell is other people” - not in an ugh mondays kind of way, but in a “you can’t be happy around others because they’re always judging you adn you’re always aware of it” way. I think that can be good (I think that social media has eroded a lot of shame around antisocial attitudes), but also bad. For instance, in Jordan, family planning services report less takeup when a woman is with her mother in law than when she is alone - the woman feels like she must, in some sense, perform for the other woman.

One of the more defining attributes of conservative social groups is that they’re very closely knit and very low trust - meaning that people scrutinize each other all the time, and everyone fears judgment. For example, Saudi husbands are actually overall fairly supportive of their wives working, but voice opposition in order to secure the approval of other men, while women are overall misinformed about their labor market outlook and the desires and aspirations of other women. Qataris express concern for women working, particularly in close proximity to men, mainly as a product of other people’s opinions of them. And in the US, single and “taken” women answer surveys about ambition similarly in private, but differently (more conservatively) in public, in order to not scare away potential boyfriends - it is social norms, not prejudice, are holding women back.

For example, it could be noted that promoting women via quotas is not especially productive at increasing representation of women leadership. In fact, women who are promoted to managerial positions are less likely to promote other women because 1) their decisiosn are more closely scrutinized, and 2) there is a concern that they would unfairly prefer women over men, resulting in a weird reverse sexism. And high-level corporate teams are substantially less likely to incorporate a second woman after the first one is hired, but the likelihood increases and then plateaus after subsequent women are hired. This is especially true if the women may be perceived to be “diversity hires” or meeting some sort of quota, even if no such quota actually exists.

In fact, sexism is currently mainly propagated via unspoken norms that disproportionately affect women, and not via explicit “get back in the kitchen” discrimination. Being a manager means working long hours and never being off the clock - hard if you’re the one looking after your kids. Just like being a scientist = being a man, to scientists at least. These unspoken norms are a massive factor in hiring decisions: for example, firms that hire Indian workers tend to discriminate by caste a lot - not by refusing to interview “untouchables” (that is literally the caste name), but by focusing on mushy intangible cultural fit attributes that just so happen to disporportionately match priest caste individuals. Generally speaking, people who hire based on “cultural fit” tend to just reiterate their own biases, which can explain some evidence that there is no racial discrimination at some stages of the hiring process - if final hiring decisions are made based on “compatibility” concerns, then perhaps they just don’t choose applicants who are Black for “company culture reasons”.

This “social conformity bias” is very powerful, and very important. In the same way that people who make unique or unconventional fashion choices are more innovative, women who conform to gender norms more tightly tend to have lower earnings and labor force participation, and vice versa. There are, in both cases, spillovers - other people benefit from this harmless inconformity. The good thing is that exposure to women breaking social norms can be beneficial for others: women sent down to work to Chinese areas with more progressive values had higher labor force participation, and men exposed to talented women learn to appreciate their contributions - such as with federal workers during the First World War.

Conclusion

So, who was right? Well, neither actually, and both actually. While it is true that explicit discrimination against women is fairly low now, and that even back in the day it wasn’t as overt as one might think (it could also help that my grandma is from a prominent family in a very small town), it still existed, and women still face many obstacles today. However, the obstacles are largely about unconscious biases and unspoken norms - and not about the kind of stuff that is prevalent in, say, Mad Men. People’s perception of gender norms is largely internalized and largely reinforced socially, and challenging those norms is both hard and important. To finish out with another family example, one of my grandma’s sisters was an accountant and ran the family firm - and my mom always emphasizes how she insisted on wearing skirts and not pants at the office, so her (male) employees would always be aware that a woman was the boss.

1. By the way this is about María Estela Martínez de Perón, who was the country’s first female president and succeeded her late husband (Juan Perón) as president in 1974. Her term saw a striking escalation of political violence and human rights violations, coupled with a complete implosion of the economy and no meaningful accomplishments beyond those.